This repository contains scripts used to create machine learning models to predict runway surface conditions from weather data and runway reports. The files contains what you need to perform a classification (slippery / non-slippery) and a regression on the tire-to-road friction coefficient, as well as explaniable AI by calculating SHAP values for the predictions.
The scripts are used for the calculations done in this paper: https://www.sciencedirect.com/science/article/pii/S0165232X22000751.
The finished trained and tuned regression model and classification model are availalbe as JSON files.
Please reach out to alisedm@math.uio.no if you have any questions.